| 引用本文: | 舒 涛,路昊天,曹景轩,等.基于混沌理论的波密气象站降雨量预测方法研究[J].灌溉排水学报,0,():-. |
| SHUTAO,LUHAOTIAN,CAOJINGXUAN,et al.基于混沌理论的波密气象站降雨量预测方法研究[J].灌溉排水学报,0,():-. |
|
| 摘要: |
| 为了得到精确度较高的月降雨量预测值，本文首先利用C-C关联积分法来确定波密站月降雨量非线性系统的延迟时间?和嵌入维数m，再对月降雨量时间序列进行相空间重构，并利用小数据量法求取Lyapunov指数来判断月降雨量时间序列的混沌特征，然后构建Volterra模型分别进行短期5年和长期15年降雨量预测，将其预测值与小波预测模型和SVR预测模型的预测值对比，最后对Volterra短期预测模型进行叠加预测误差分析和模型推广分析。研究表明:Volterra模型对混沌特征明显的月降雨量进行短期预测时，其MAPE和EC分为4.04%和0.941，相比小波和SVR模型来说具有较高的预测精度，同时叠加预测误差较小，其MAPE为7.657%，EC为0.894；而在长期预测时，该模型预测精度不如SVR模型；同时Volterra模型对混沌特征弱的月降雨量进行短期预测时，其模型预测效果并不理想，MAPE为54.855%，EC仅为0.566。故该方法能够较准确的预测短期混沌降雨量。 |
| 关键词: 混沌理论;相空间重构;Lyapunov指数;Volterra滤波器;降雨量预测 |
| DOI： |
| 分类号:P457;P208 |
| 基金项目:广西研究生教育创新计划项目(YCSW2021203)；大学生创新实验项目(202010694008)；大学生创新实验项目(2020XCX011) |
|
| Research on Precipitation Prediction Method of Bomi Meteorological Station Based on Chaos Theory |
|
SHUTAO1, LUHAOTIAN2, CAOJINGXUAN3, YETANGJIN4, TAOWEI3, FURunYI3, LIHAO5
|
|
1.College of Marine Geosciences, Ocean University of China;2.Guilin University of Technology;3.College of Engineering, Tibet University;4.Department of Construction Engineering, Dalian University of Technology;5.Key Laboratory of Mountain Hazards and Earth Surface Process, Chinese Academy of Sciences
|
| Abstract: |
| In order to obtain High-precision monthly rainfall predictive value, In this paper, the C-C correlation integral method is used to determine the delay time ? and embedding dimension m of the nonlinear system of monthly rainfall at the Bomi station. Moreover, the monthly rainfall time series is reconstructed phase space, and the Lyapunov exponent is obtained by using the small data sets method to determine the chaotic characteristics of the monthly rainfall time series. Afterward, the Volterra model is constructed to respectively predict the short-term 5-year and long-term 15-year rainfall. The predicted values are compared with the predicted values of the wavelet prediction model and the SVR prediction model. Finally, the superposition prediction error analysis and model promotion analysis of the Volterra short-term prediction model is carried out. The results show that the MAPE and EC values of the Volterra model respectively are 4.04 % and 0.941 for the short-term prediction of monthly rainfall with obvious chaotic characteristics. Compared with the wavelet and SVR models, the Volterra model has higher prediction accuracy, and the superposition prediction error is smaller, the MAPE and EC values of the Volterra model are 7.657% and 0.894 respectively. Moreover, in the long-term prediction, the prediction accuracy of this model is not as good as the SVR model. At the same time, when the Volterra model is used for short-term prediction of monthly rainfall with weak chaotic characteristics, the prediction effect of the model is not ideal, with a MAPE value of 54.855 % and EC value of 0.566. Therefore, this method can accurately predict short-term chaotic rainfall. |
| Key words: Chaos Theory; Phase Space Reconstruction; Lyapunov Index; Volterra Filter; Rainfall Prediction? |
